Type 2 diabetes mellitus (DM) is a major source of morbidity and mortality worldwide. Cases of DM are expected to rise by 72% through 2025, affecting 325 million people across all nations and income groups. It is increasingly recognized that there is phenotypic heterogeneity among individuals who develop DM. One subgroup that has attracted particular attention in recent years is the lean diabetes group, e.g. individuals with DM in the absence of obesity. Lean individuals with DM are at substantially higher risk of mortality and other complications than obese individuals with DM. However, little is known about the factors that promote DM in the absence of obesity, or the mechanisms underlying the worse outcomes in this subset. Asians are particularly susceptible to developing lean DM. Approximately half of Asians who develop DM are considered normal weight (BMI less than 25 kg/m2). This propensity is independent of country of residence, e.g. it affects Asians living in th U.S. as well as in Asian countries. Indeed, studies in the U.S. indicate very high rates of DM among Asian-Americans. The pathogenesis of DM reflects a complex interplay of genetic, dietary, and environmental exposures affecting multiple pathways. One approach to understanding the activity in many metabolic pathways at once is metabolomics profiling. Metabolomics refers to the systematic analysis of metabolites in a biological specimen, such as plasma. Combining biomarker and phenotypic data in human populations provides a rich opportunity to identify the biochemical signatures of metabolic diseases, which can enhance biological understanding as well as yield tools for disease screening. Metabolomics data from non-European cohorts, and particularly Asian cohorts, are sparse. The differences in the epidemiology of DM across racial/ethnic groups suggest the possibility of pathophysiological differences. Recently, in a preliminary study in the Shanghai Women's Health Study (SWHS), we found evidence that Chinese women had some risk markers for DM that were distinct from those described in European populations. Thus, we propose to expand considerably on our work in European populations and our pilot studies in the SWHS, by performing comprehensive metabolomics profiling in 2 Chinese cohorts to identify metabolites that are associated with incident DM. Our aims are (1) to identify metabolites that associate with incident DM in Chinese individuals enrolled in the SWHS and Shanghai Men's Health Study (SMHS); and (2) to replicate the association of metabolites with incident DM in a separate case-cohort study. We will also assess whether addition of metabolites improves the ability to predict risk of DM in Asians, compared with risk factors alone, and whether the metabolite predictors are associated with mortality and cardiovascular events. Our team is uniquely positioned to conduct these studies, given our expertise in applying metabolomics in large cohorts, our access to other metabolomics datasets for comparison and validation, and our experience with chronic disease epidemiology in China.